Looking forward to your support on the Hydra proposal. Please reach out if you have any questions.

Most of the L2 Scalability proposal's ask (80%) is for Hydra, so it's worth revisiting what Hydra is and what kinds of use cases it enables for Cardano. Hydra is Cardano's Layer 2 state channel scaling solution. It lets participants open off-chain state channels where transactions settle in milliseconds with zero fees, while inheriting the full security of Cardano L1 as a backstop. USPs: Sub-second finality. No waiting for block confirmations inside a head. Zero fees. Protocol-level, not a subsidy. High TPS. Process thousands of TPS inside the head. Isomorphic. Full smart contract support inside the head, same model as L1. L1-anchored security. Disputes always settle on Cardano mainnet. Funds cannot be taken without unanimous consent. Privacy. Intermediate states are known only to head participants. Trust assumptions and use cases Hydra operates in two modes. Conflating them leads to confusion. Direct setup: participants run their own hydra-nodes. Every participant co-signs every snapshot. The head state cannot advance without your signature. No operator trust at all. This would ideal for institutional market makers, financial actors, and autonomous agents who already run infrastructure. Hydra is not meant for retail use in the direct setup mode. It can, though, in the delegated setup mentioned below. Direct-setup use cases with high ROI: Institutional OTC settlement. Market makers in a Hydra head arbitraging across DEXes by OTC trading between each other: sourcing liquidity, rebalancing inventory, executing liquidation-to-USD flows, all at sub-second finality with zero fees. In mature DeFi markets, institutional traders dominate trading volume even though retail dominates TVL. That volume is Hydra's market. Agent-to-agent commerce. In Masumi's model, buyer and seller agents each run their own hydra-node. They are the participants. They sign every snapshot. This is direct setup because agents are software: running a hydra-node is operationally identical to running an API server. Circle launched Nanopayments in March 2026 for this exact market: off-chain aggregation, batched L1 settlement, the same architecture as Hydra. But Nanopayments is Circle-dependent, USDC-only, and payment-only. Masumi on Hydra is a full commerce layer: escrow, dispute resolution, programmable service contracts, continuous micropayment streams. x402 and Nanopayments let an agent pay for an API call. Masumi lets agents run actual commerce. Citi projects hundreds of billions in agent-driven transaction value by 2030. Cardano has a first-mover advantage here. B2B payment rails. Known counterparties, contractual relationship, both run infrastructure. Net out flows throughout the day, settle to L1 once. Delegated setup: operators run the head, users transact through an application layer. Operators control execution, but unanimous consensus means safety is preserved as long as one honest operator remains. This works where trust points already exist and Hydra adds no custodial risk on top: Prediction markets already trust a resolver to settle outcomes. If the operator set overlaps with the resolver set, Hydra adds zero marginal trust. Polymarket's election cycle did $9 billion in volume. Perp DEXes accept federated operator trust in exchange for CEX-like performance. Hyperliquid clears around $5 billion daily on a 25-validator BFT chain. Delta DeFi is the closest Cardano analog, with L1 settlement as a stronger backstop. RWA platforms are federated by regulatory necessity: custodians, auditors, KYC'd participants. Hydra absorbs trust that already exists. BlackRock's BUIDL passed $2.5 billion in 2025, tokenized treasuries crossed $7 billion, BCG projects $16 trillion by 2030. Other use cases include gaming, point of sales and general information processing like voting applications. The pattern across all of these: trust already exists in the application layer. Hydra absorbs it into a structured protocol with cryptographic settlement guarantees rather than introducing new trust. The realistic comparison is not 'trustless versus custodial' but 'fully centralized service versus federated head with L1-anchored settlement.' Most of the L2 landscape including Optimism, Base, and Arbitrum runs on trusted sequencers today, and Hydra's delegated setup provides stronger custody guarantees through unanimous-consent settlement. The proposal includes Hydra reference implementations for such patterns, so teams have a working starting point rather than building from scratch. Lowering deployment friction is key to adoption. Protocol maturity Hydra has been in R&D for a while. The past year changed the game: Hydra Doom: massive success, huge buzz for Cardano Glacier Drop: 30M+ users, 8 blockchains. Most rigorous stress test any Cardano L2 has faced Delta DeFi: production perp DEX on Hydra Masumi: agent-to-agent commerce, live VTech Labs: building applications on top Midgard: using Hydra for fast withdrawals Intersect voting: runs on Hydra Hydra Vending Machine Hydra now has a growing production base. Early adopters uncovered issues. Many fixed have been fixed. The difference now: we have real feedback loops from actual users which is helping production hardening. We did not have that before. The Hydra team has delivered: v1.3: very stable and performant, addressing issues from production users v2 alpha: out now, with directly opening heads (simpler workflow, lower costs) Partial fanout: releasing soon, ahead of milestone deadline Hydra is Cardano's most mature, stress-tested scaling solution. Actual users depend on it in production. The hardening work in this proposal closes the remaining gap those users have surfaced and ensures they get the support they need as users.

Hydra v2 (alpha) is here! This release simplifies the Hydra Head protocol by removing the commit phase entirely: heads now open directly, and funds are added incrementally via deposits. This eliminates collectCom and abort transactions, resolves the long-standing "non-abortable head" issue, and reduces lifecycle costs for most use cases. More details: github.com/cardano-scaling/h… Please try it out and give us feedback!